@promptbook/legacy-documents 0.71.0-18 → 0.71.0-20
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +0 -6
- package/esm/index.es.js +80 -86
- package/esm/index.es.js.map +1 -1
- package/esm/typings/src/_packages/core.index.d.ts +22 -18
- package/esm/typings/src/_packages/types.index.d.ts +2 -2
- package/esm/typings/src/collection/PipelineCollection.d.ts +1 -1
- package/esm/typings/src/collection/SimplePipelineCollection.d.ts +2 -2
- package/esm/typings/src/collection/collectionToJson.d.ts +1 -1
- package/esm/typings/src/collection/constructors/createCollectionFromJson.d.ts +1 -1
- package/esm/typings/src/collection/constructors/createCollectionFromPromise.d.ts +1 -1
- package/esm/typings/src/commands/_common/types/CommandParser.d.ts +5 -5
- package/esm/typings/src/config.d.ts +21 -14
- package/esm/typings/src/execution/EmbeddingVector.d.ts +1 -1
- package/esm/typings/src/execution/LlmExecutionTools.d.ts +1 -1
- package/esm/typings/src/execution/PipelineExecutorResult.d.ts +2 -2
- package/esm/typings/src/execution/assertsExecutionSuccessful.d.ts +1 -0
- package/esm/typings/src/execution/createPipelineExecutor/00-CreatePipelineExecutorOptions.d.ts +29 -6
- package/esm/typings/src/execution/createPipelineExecutor/10-executePipeline.d.ts +2 -11
- package/esm/typings/src/execution/createPipelineExecutor/20-executeTemplate.d.ts +4 -13
- package/esm/typings/src/execution/createPipelineExecutor/40-executeAttempts.d.ts +9 -14
- package/esm/typings/src/execution/utils/addUsage.d.ts +1 -1
- package/esm/typings/src/execution/utils/forEachAsync.d.ts +1 -1
- package/esm/typings/src/formats/_common/FormatDefinition.d.ts +2 -2
- package/esm/typings/src/formats/_common/FormatSubvalueDefinition.d.ts +2 -2
- package/esm/typings/src/llm-providers/_common/register/$provideLlmToolsForCli.d.ts +2 -2
- package/esm/typings/src/llm-providers/_common/register/LlmToolsConfiguration.d.ts +1 -1
- package/esm/typings/src/llm-providers/_common/utils/cache/CacheLlmToolsOptions.d.ts +4 -1
- package/esm/typings/src/llm-providers/anthropic-claude/AnthropicClaudeExecutionTools.d.ts +1 -1
- package/esm/typings/src/llm-providers/anthropic-claude/anthropic-claude-models.d.ts +1 -1
- package/esm/typings/src/llm-providers/azure-openai/AzureOpenAiExecutionTools.d.ts +1 -1
- package/esm/typings/src/llm-providers/mocked/$fakeTextToExpectations.d.ts +1 -1
- package/esm/typings/src/llm-providers/mocked/MockedEchoLlmExecutionTools.d.ts +1 -1
- package/esm/typings/src/llm-providers/mocked/MockedFackedLlmExecutionTools.d.ts +1 -1
- package/esm/typings/src/llm-providers/multiple/MultipleLlmExecutionTools.d.ts +3 -3
- package/esm/typings/src/llm-providers/multiple/joinLlmExecutionTools.d.ts +1 -1
- package/esm/typings/src/llm-providers/openai/OpenAiExecutionTools.d.ts +1 -1
- package/esm/typings/src/llm-providers/openai/openai-models.d.ts +1 -1
- package/esm/typings/src/llm-providers/remote/RemoteLlmExecutionTools.d.ts +1 -1
- package/esm/typings/src/llm-providers/remote/interfaces/PromptbookServer_ListModels_Response.d.ts +1 -1
- package/esm/typings/src/prepare/PrepareAndScrapeOptions.d.ts +4 -4
- package/esm/typings/src/prepare/prepareTemplates.d.ts +1 -1
- package/esm/typings/src/scrapers/_common/Scraper.d.ts +1 -5
- package/esm/typings/src/scrapers/_common/prepareKnowledgePieces.d.ts +1 -1
- package/esm/typings/src/scrapers/_common/register/$provideScrapersForBrowser.d.ts +1 -1
- package/esm/typings/src/scrapers/_common/register/$provideScrapersForNode.d.ts +1 -1
- package/esm/typings/src/scrapers/_common/register/ScraperAndConverterMetadata.d.ts +2 -2
- package/esm/typings/src/scrapers/_common/utils/getScraperIntermediateSource.d.ts +2 -2
- package/esm/typings/src/scrapers/_common/utils/makeKnowledgeSourceHandler.d.ts +1 -1
- package/esm/typings/src/scrapers/document/DocumentScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/document/createDocumentScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/document/register-metadata.d.ts +1 -1
- package/esm/typings/src/scrapers/document-legacy/LegacyDocumentScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/document-legacy/createLegacyDocumentScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/document-legacy/register-metadata.d.ts +1 -1
- package/esm/typings/src/scrapers/markdown/MarkdownScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/markdown/createMarkdownScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/markdown/register-metadata.d.ts +1 -1
- package/esm/typings/src/scrapers/pdf/PdfScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/pdf/createPdfScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/pdf/register-metadata.d.ts +1 -1
- package/esm/typings/src/scrapers/website/WebsiteScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/website/createWebsiteScraper.d.ts +1 -1
- package/esm/typings/src/scrapers/website/register-metadata.d.ts +1 -1
- package/esm/typings/src/scripting/javascript/utils/preserve.d.ts +1 -1
- package/esm/typings/src/storage/file-cache-storage/utils/nameToSubfolderPath.d.ts +1 -1
- package/esm/typings/src/types/Arrayable.d.ts +1 -1
- package/esm/typings/src/types/IntermediateFilesStrategy.d.ts +7 -0
- package/esm/typings/src/types/PipelineJson/KnowledgePieceJson.d.ts +4 -4
- package/esm/typings/src/types/PipelineJson/KnowledgeSourceJson.d.ts +1 -1
- package/esm/typings/src/types/PipelineJson/PersonaJson.d.ts +1 -1
- package/esm/typings/src/types/Prompt.d.ts +1 -1
- package/esm/typings/src/types/execution-report/ExecutionReportJson.d.ts +1 -1
- package/esm/typings/src/utils/$Register.d.ts +1 -1
- package/esm/typings/src/utils/FromtoItems.d.ts +1 -1
- package/esm/typings/src/utils/arrayableToArray.d.ts +1 -1
- package/esm/typings/src/utils/emojis.d.ts +1 -1
- package/esm/typings/src/utils/expectation-counters/countSentences.d.ts +1 -1
- package/esm/typings/src/utils/markdown/extractAllBlocksFromMarkdown.d.ts +1 -1
- package/esm/typings/src/utils/markdown/splitMarkdownIntoSections.d.ts +1 -1
- package/esm/typings/src/utils/organization/TODO_USE.d.ts +1 -1
- package/esm/typings/src/utils/organization/keepUnused.d.ts +1 -1
- package/esm/typings/src/utils/random/$randomSeed.d.ts +1 -1
- package/esm/typings/src/utils/sets/intersection.d.ts +1 -1
- package/esm/typings/src/utils/sets/union.d.ts +1 -1
- package/package.json +2 -2
- package/umd/index.umd.js +80 -86
- package/umd/index.umd.js.map +1 -1
- package/esm/typings/src/execution/createPipelineExecutor/00-CreatePipelineExecutorSettings.d.ts +0 -29
package/README.md
CHANGED
|
@@ -20,12 +20,6 @@ Build responsible, controlled and transparent applications on top of LLM models!
|
|
|
20
20
|
|
|
21
21
|
- ✨ **Support of [OpenAI o1 model](https://openai.com/o1/)**
|
|
22
22
|
|
|
23
|
-
<blockquote style="color: #ff8811">
|
|
24
|
-
<b>⚠ Warning:</b> This is a pre-release version of the library. It is not yet ready for production use. Please look at <a href="https://www.npmjs.com/package/@promptbook/core?activeTab=versions">latest stable release</a>.
|
|
25
|
-
</blockquote>
|
|
26
|
-
|
|
27
|
-
|
|
28
|
-
|
|
29
23
|
<blockquote style="color: #ff8811">
|
|
30
24
|
<b>⚠ Warning:</b> This is a pre-release version of the library. It is not yet ready for production use. Please look at <a href="https://www.npmjs.com/package/@promptbook/core?activeTab=versions">latest stable release</a>.
|
|
31
25
|
</blockquote>
|
package/esm/index.es.js
CHANGED
|
@@ -15,7 +15,7 @@ import { unparse, parse } from 'papaparse';
|
|
|
15
15
|
/**
|
|
16
16
|
* The version of the Promptbook library
|
|
17
17
|
*/
|
|
18
|
-
var PROMPTBOOK_VERSION = '0.71.0-
|
|
18
|
+
var PROMPTBOOK_VERSION = '0.71.0-19';
|
|
19
19
|
// TODO: [main] !!!! List here all the versions and annotate + put into script
|
|
20
20
|
|
|
21
21
|
/*! *****************************************************************************
|
|
@@ -363,18 +363,25 @@ var LOOP_LIMIT = 1000;
|
|
|
363
363
|
* @private within the repository - too low-level in comparison with other `MAX_...`
|
|
364
364
|
*/
|
|
365
365
|
var IMMEDIATE_TIME = 10;
|
|
366
|
+
/**
|
|
367
|
+
* Strategy for caching the intermediate results for knowledge sources
|
|
368
|
+
*
|
|
369
|
+
* @public exported from `@promptbook/core`
|
|
370
|
+
*/
|
|
371
|
+
var DEFAULT_INTERMEDIATE_FILES_STRATEGY = 'HIDE_AND_KEEP';
|
|
372
|
+
// <- TODO: [😡] Change to 'VISIBLE'
|
|
366
373
|
/**
|
|
367
374
|
* The maximum number of (LLM) tasks running in parallel
|
|
368
375
|
*
|
|
369
376
|
* @public exported from `@promptbook/core`
|
|
370
377
|
*/
|
|
371
|
-
var
|
|
378
|
+
var DEFAULT_MAX_PARALLEL_COUNT = 5; // <- TODO: [🤹♂️]
|
|
372
379
|
/**
|
|
373
380
|
* The maximum number of attempts to execute LLM task before giving up
|
|
374
381
|
*
|
|
375
382
|
* @public exported from `@promptbook/core`
|
|
376
383
|
*/
|
|
377
|
-
var
|
|
384
|
+
var DEFAULT_MAX_EXECUTION_ATTEMPTS = 3; // <- TODO: [🤹♂️]
|
|
378
385
|
/**
|
|
379
386
|
* Where to store the scrape cache
|
|
380
387
|
*
|
|
@@ -382,7 +389,7 @@ var MAX_EXECUTION_ATTEMPTS = 3; // <- TODO: [🤹♂️]
|
|
|
382
389
|
*
|
|
383
390
|
* @public exported from `@promptbook/core`
|
|
384
391
|
*/
|
|
385
|
-
var
|
|
392
|
+
var DEFAULT_SCRAPE_CACHE_DIRNAME = '/.promptbook/scrape-cache';
|
|
386
393
|
/**
|
|
387
394
|
* Nonce which is used for replacing things in strings
|
|
388
395
|
*
|
|
@@ -434,7 +441,7 @@ var DEFAULT_CSV_SETTINGS = Object.freeze({
|
|
|
434
441
|
*
|
|
435
442
|
* @public exported from `@promptbook/core`
|
|
436
443
|
*/
|
|
437
|
-
var
|
|
444
|
+
var DEFAULT_IS_VERBOSE = false;
|
|
438
445
|
/**
|
|
439
446
|
* @@@
|
|
440
447
|
*
|
|
@@ -1150,12 +1157,12 @@ function TODO_USE() {
|
|
|
1150
1157
|
*/
|
|
1151
1158
|
function getScraperIntermediateSource(source, options) {
|
|
1152
1159
|
return __awaiter(this, void 0, void 0, function () {
|
|
1153
|
-
var sourceFilename, url, rootDirname, cacheDirname,
|
|
1160
|
+
var sourceFilename, url, rootDirname, cacheDirname, intermediateFilesStrategy, extension, isVerbose, hash, semanticName, pieces, name, cacheFilename, isDestroyed, fileHandler;
|
|
1154
1161
|
return __generator(this, function (_a) {
|
|
1155
1162
|
switch (_a.label) {
|
|
1156
1163
|
case 0:
|
|
1157
1164
|
sourceFilename = source.filename, url = source.url;
|
|
1158
|
-
rootDirname = options.rootDirname, cacheDirname = options.cacheDirname,
|
|
1165
|
+
rootDirname = options.rootDirname, cacheDirname = options.cacheDirname, intermediateFilesStrategy = options.intermediateFilesStrategy, extension = options.extension, isVerbose = options.isVerbose;
|
|
1159
1166
|
hash = SHA256(
|
|
1160
1167
|
// <- TODO: [🥬] Encapsulate sha256 to some private utility function
|
|
1161
1168
|
hexEncoder.parse(sourceFilename || url || 'untitled'))
|
|
@@ -1165,7 +1172,7 @@ function getScraperIntermediateSource(source, options) {
|
|
|
1165
1172
|
pieces = ['intermediate', semanticName, hash].filter(function (piece) { return piece !== ''; });
|
|
1166
1173
|
name = pieces.join('-').split('--').join('-');
|
|
1167
1174
|
// <- TODO: Use MAX_FILENAME_LENGTH
|
|
1168
|
-
TODO_USE(rootDirname); // <- TODO:
|
|
1175
|
+
TODO_USE(rootDirname); // <- TODO: [😡]
|
|
1169
1176
|
cacheFilename = join.apply(void 0, __spreadArray(__spreadArray([process.cwd(),
|
|
1170
1177
|
cacheDirname], __read(nameToSubfolderPath(hash /* <- TODO: [🎎] Maybe add some SHA256 prefix */)), false), [name], false)).split('\\')
|
|
1171
1178
|
.join('/') +
|
|
@@ -1185,7 +1192,7 @@ function getScraperIntermediateSource(source, options) {
|
|
|
1185
1192
|
return __generator(this, function (_a) {
|
|
1186
1193
|
switch (_a.label) {
|
|
1187
1194
|
case 0:
|
|
1188
|
-
if (!
|
|
1195
|
+
if (!(intermediateFilesStrategy === 'HIDE_AND_CLEAN')) return [3 /*break*/, 2];
|
|
1189
1196
|
if (isVerbose) {
|
|
1190
1197
|
console.info('legacyDocumentScraper: Clening cache');
|
|
1191
1198
|
}
|
|
@@ -1209,7 +1216,7 @@ function getScraperIntermediateSource(source, options) {
|
|
|
1209
1216
|
/**
|
|
1210
1217
|
* Note: Not using `FileCacheStorage` for two reasons:
|
|
1211
1218
|
* 1) Need to store more than serialized JSONs
|
|
1212
|
-
* 2) Need to switch between a `rootDirname` and `cacheDirname` <- TODO:
|
|
1219
|
+
* 2) Need to switch between a `rootDirname` and `cacheDirname` <- TODO: [😡]
|
|
1213
1220
|
* TODO: [🐱🐉][🧠] Make some smart crop
|
|
1214
1221
|
* Note: [🟢] Code in this file should never be never released in packages that could be imported into browser environment
|
|
1215
1222
|
*/
|
|
@@ -1274,7 +1281,6 @@ function pipelineJsonToString(pipelineJson) {
|
|
|
1274
1281
|
pipelineString += '\n\n';
|
|
1275
1282
|
pipelineString += description;
|
|
1276
1283
|
}
|
|
1277
|
-
// TODO:> const commands: Array<Command>
|
|
1278
1284
|
var commands = [];
|
|
1279
1285
|
if (pipelineUrl) {
|
|
1280
1286
|
commands.push("PIPELINE URL ".concat(pipelineUrl));
|
|
@@ -1330,7 +1336,6 @@ function pipelineJsonToString(pipelineJson) {
|
|
|
1330
1336
|
pipelineString += '\n\n';
|
|
1331
1337
|
pipelineString += description_1;
|
|
1332
1338
|
}
|
|
1333
|
-
// TODO:> const commands: Array<Command>
|
|
1334
1339
|
var commands_1 = [];
|
|
1335
1340
|
var contentLanguage = 'text';
|
|
1336
1341
|
if (templateType === 'PROMPT_TEMPLATE') {
|
|
@@ -2215,6 +2220,7 @@ function assertsExecutionSuccessful(executionResult) {
|
|
|
2215
2220
|
}
|
|
2216
2221
|
}
|
|
2217
2222
|
/**
|
|
2223
|
+
* TODO: [🐚] This function should be removed OR changed OR be completely rewritten
|
|
2218
2224
|
* TODO: [🧠] Can this return type be better typed than void
|
|
2219
2225
|
*/
|
|
2220
2226
|
|
|
@@ -2875,7 +2881,7 @@ function preparePersona(personaDescription, tools, options) {
|
|
|
2875
2881
|
return __generator(this, function (_d) {
|
|
2876
2882
|
switch (_d.label) {
|
|
2877
2883
|
case 0:
|
|
2878
|
-
_a = options.isVerbose, isVerbose = _a === void 0 ?
|
|
2884
|
+
_a = options.isVerbose, isVerbose = _a === void 0 ? DEFAULT_IS_VERBOSE : _a;
|
|
2879
2885
|
if (tools === undefined || tools.llm === undefined) {
|
|
2880
2886
|
throw new MissingToolsError('LLM tools are required for preparing persona');
|
|
2881
2887
|
}
|
|
@@ -3041,7 +3047,7 @@ var $Register = /** @class */ (function () {
|
|
|
3041
3047
|
this.storage = globalScope[storageName];
|
|
3042
3048
|
}
|
|
3043
3049
|
$Register.prototype.list = function () {
|
|
3044
|
-
// <- TODO: ReadonlyDeep<
|
|
3050
|
+
// <- TODO: ReadonlyDeep<ReadonlyArray<TRegistered>>
|
|
3045
3051
|
return this.storage;
|
|
3046
3052
|
};
|
|
3047
3053
|
$Register.prototype.register = function (registered) {
|
|
@@ -3228,7 +3234,7 @@ function extensionToMimeType(value) {
|
|
|
3228
3234
|
/**
|
|
3229
3235
|
* @@@
|
|
3230
3236
|
*
|
|
3231
|
-
* @
|
|
3237
|
+
* @public exported from `@promptbook/core`
|
|
3232
3238
|
*/
|
|
3233
3239
|
function makeKnowledgeSourceHandler(knowledgeSource, tools, options) {
|
|
3234
3240
|
var _a;
|
|
@@ -3241,7 +3247,7 @@ function makeKnowledgeSourceHandler(knowledgeSource, tools, options) {
|
|
|
3241
3247
|
case 0:
|
|
3242
3248
|
sourceContent = knowledgeSource.sourceContent;
|
|
3243
3249
|
name = knowledgeSource.name;
|
|
3244
|
-
_b = options || {}, _c = _b.rootDirname, rootDirname = _c === void 0 ? null : _c, _d = _b.isVerbose, isVerbose = _d === void 0 ?
|
|
3250
|
+
_b = options || {}, _c = _b.rootDirname, rootDirname = _c === void 0 ? null : _c, _d = _b.isVerbose, isVerbose = _d === void 0 ? DEFAULT_IS_VERBOSE : _d;
|
|
3245
3251
|
TODO_USE(isVerbose);
|
|
3246
3252
|
if (!name) {
|
|
3247
3253
|
name = sourceContentToName(sourceContent);
|
|
@@ -3394,7 +3400,7 @@ function prepareKnowledgePieces(knowledgeSources, tools, options) {
|
|
|
3394
3400
|
return __generator(this, function (_c) {
|
|
3395
3401
|
switch (_c.label) {
|
|
3396
3402
|
case 0:
|
|
3397
|
-
_a = options.maxParallelCount, maxParallelCount = _a === void 0 ?
|
|
3403
|
+
_a = options.maxParallelCount, maxParallelCount = _a === void 0 ? DEFAULT_MAX_PARALLEL_COUNT : _a, rootDirname = options.rootDirname, _b = options.isVerbose, isVerbose = _b === void 0 ? DEFAULT_IS_VERBOSE : _b;
|
|
3398
3404
|
knowledgePreparedUnflatten = new Array(knowledgeSources.length);
|
|
3399
3405
|
return [4 /*yield*/, forEachAsync(knowledgeSources, { maxParallelCount: maxParallelCount }, function (knowledgeSource, index) { return __awaiter(_this, void 0, void 0, function () {
|
|
3400
3406
|
var partialPieces, sourceHandler, _a, _b, scraper, partialPiecesUnchecked, e_1_1, pieces;
|
|
@@ -3423,7 +3429,8 @@ function prepareKnowledgePieces(knowledgeSources, tools, options) {
|
|
|
3423
3429
|
case 4:
|
|
3424
3430
|
partialPiecesUnchecked = _d.sent();
|
|
3425
3431
|
if (partialPiecesUnchecked !== null) {
|
|
3426
|
-
partialPieces = partialPiecesUnchecked;
|
|
3432
|
+
partialPieces = __spreadArray([], __read(partialPiecesUnchecked), false);
|
|
3433
|
+
// <- TODO: [🪓] Here should be no need for spreading new array, just `partialPieces = partialPiecesUnchecked`
|
|
3427
3434
|
return [3 /*break*/, 6];
|
|
3428
3435
|
}
|
|
3429
3436
|
_d.label = 5;
|
|
@@ -3471,7 +3478,7 @@ TODO: [🧊] This is how it can look in future
|
|
|
3471
3478
|
> /**
|
|
3472
3479
|
> * Unprepared knowledge
|
|
3473
3480
|
> * /
|
|
3474
|
-
> readonly knowledgeSources:
|
|
3481
|
+
> readonly knowledgeSources: ReadonlyArray<KnowledgeSourceJson>;
|
|
3475
3482
|
> };
|
|
3476
3483
|
>
|
|
3477
3484
|
> export async function prepareKnowledgePieces(
|
|
@@ -3529,7 +3536,7 @@ function prepareTemplates(pipeline, tools, options) {
|
|
|
3529
3536
|
return __generator(this, function (_b) {
|
|
3530
3537
|
switch (_b.label) {
|
|
3531
3538
|
case 0:
|
|
3532
|
-
_a = options.maxParallelCount, maxParallelCount = _a === void 0 ?
|
|
3539
|
+
_a = options.maxParallelCount, maxParallelCount = _a === void 0 ? DEFAULT_MAX_PARALLEL_COUNT : _a;
|
|
3533
3540
|
templates = pipeline.templates, parameters = pipeline.parameters, knowledgePiecesCount = pipeline.knowledgePiecesCount;
|
|
3534
3541
|
// TODO: [main] !!!!! Apply samples to each template (if missing and is for the template defined)
|
|
3535
3542
|
TODO_USE(parameters);
|
|
@@ -3591,7 +3598,7 @@ function preparePipeline(pipeline, tools, options) {
|
|
|
3591
3598
|
if (isPipelinePrepared(pipeline)) {
|
|
3592
3599
|
return [2 /*return*/, pipeline];
|
|
3593
3600
|
}
|
|
3594
|
-
rootDirname = options.rootDirname, _a = options.maxParallelCount, maxParallelCount = _a === void 0 ?
|
|
3601
|
+
rootDirname = options.rootDirname, _a = options.maxParallelCount, maxParallelCount = _a === void 0 ? DEFAULT_MAX_PARALLEL_COUNT : _a, _b = options.isVerbose, isVerbose = _b === void 0 ? DEFAULT_IS_VERBOSE : _b;
|
|
3595
3602
|
parameters = pipeline.parameters, templates = pipeline.templates, knowledgeSources = pipeline.knowledgeSources, personas = pipeline.personas;
|
|
3596
3603
|
if (tools === undefined || tools.llm === undefined) {
|
|
3597
3604
|
throw new MissingToolsError('LLM tools are required for preparing the pipeline');
|
|
@@ -3649,7 +3656,9 @@ function preparePipeline(pipeline, tools, options) {
|
|
|
3649
3656
|
// ----- /Templates preparation -----
|
|
3650
3657
|
// Note: Count total usage
|
|
3651
3658
|
currentPreparation.usage = llmToolsWithUsage.getTotalUsage();
|
|
3652
|
-
return [2 /*return*/, $asDeeplyFrozenSerializableJson('Prepared PipelineJson', __assign(__assign({}, clonePipeline(pipeline)), { templates:
|
|
3659
|
+
return [2 /*return*/, $asDeeplyFrozenSerializableJson('Prepared PipelineJson', __assign(__assign({}, clonePipeline(pipeline)), { templates: __spreadArray([], __read(templatesPrepared), false),
|
|
3660
|
+
// <- TODO: [🪓] Here should be no need for spreading new array, just ` templates: templatesPrepared`
|
|
3661
|
+
knowledgeSources: knowledgeSourcesPrepared, knowledgePieces: knowledgePiecesPrepared, personas: preparedPersonas, preparations: __spreadArray([], __read(preparations), false) }))];
|
|
3653
3662
|
}
|
|
3654
3663
|
});
|
|
3655
3664
|
});
|
|
@@ -4620,12 +4629,11 @@ function checkExpectations(expectations, value) {
|
|
|
4620
4629
|
*/
|
|
4621
4630
|
function executeAttempts(options) {
|
|
4622
4631
|
return __awaiter(this, void 0, void 0, function () {
|
|
4623
|
-
var jokerParameterNames, priority, maxAttempts, preparedContent, parameters, template, preparedPipeline, tools,
|
|
4632
|
+
var jokerParameterNames, priority, maxAttempts, preparedContent, parameters, template, preparedPipeline, tools, $executionReport, pipelineIdentification, maxExecutionAttempts, $ongoingTemplateResult, _llms, llmTools, _loop_1, attempt, state_1;
|
|
4624
4633
|
return __generator(this, function (_a) {
|
|
4625
4634
|
switch (_a.label) {
|
|
4626
4635
|
case 0:
|
|
4627
|
-
jokerParameterNames = options.jokerParameterNames, priority = options.priority, maxAttempts = options.maxAttempts, preparedContent = options.preparedContent, parameters = options.parameters, template = options.template, preparedPipeline = options.preparedPipeline, tools = options.tools,
|
|
4628
|
-
maxExecutionAttempts = settings.maxExecutionAttempts;
|
|
4636
|
+
jokerParameterNames = options.jokerParameterNames, priority = options.priority, maxAttempts = options.maxAttempts, preparedContent = options.preparedContent, parameters = options.parameters, template = options.template, preparedPipeline = options.preparedPipeline, tools = options.tools, $executionReport = options.$executionReport, pipelineIdentification = options.pipelineIdentification, maxExecutionAttempts = options.maxExecutionAttempts;
|
|
4629
4637
|
$ongoingTemplateResult = {
|
|
4630
4638
|
$result: null,
|
|
4631
4639
|
$resultString: null,
|
|
@@ -4991,12 +4999,12 @@ function executeAttempts(options) {
|
|
|
4991
4999
|
*/
|
|
4992
5000
|
function executeFormatSubvalues(options) {
|
|
4993
5001
|
return __awaiter(this, void 0, void 0, function () {
|
|
4994
|
-
var template, jokerParameterNames, parameters, priority,
|
|
5002
|
+
var template, jokerParameterNames, parameters, priority, csvSettings, pipelineIdentification, parameterValue, formatDefinition, subvalueDefinition, formatSettings, resultString;
|
|
4995
5003
|
var _this = this;
|
|
4996
5004
|
return __generator(this, function (_a) {
|
|
4997
5005
|
switch (_a.label) {
|
|
4998
5006
|
case 0:
|
|
4999
|
-
template = options.template, jokerParameterNames = options.jokerParameterNames, parameters = options.parameters, priority = options.priority,
|
|
5007
|
+
template = options.template, jokerParameterNames = options.jokerParameterNames, parameters = options.parameters, priority = options.priority, csvSettings = options.csvSettings, pipelineIdentification = options.pipelineIdentification;
|
|
5000
5008
|
if (template.foreach === undefined) {
|
|
5001
5009
|
return [2 /*return*/, /* not await */ executeAttempts(options)];
|
|
5002
5010
|
}
|
|
@@ -5026,7 +5034,7 @@ function executeFormatSubvalues(options) {
|
|
|
5026
5034
|
.join('\n')), "\n\n [\u26F7] This should never happen because subformat name should be validated during parsing\n\n ").concat(block(pipelineIdentification), "\n "); }));
|
|
5027
5035
|
}
|
|
5028
5036
|
if (formatDefinition.formatName === 'CSV') {
|
|
5029
|
-
formatSettings =
|
|
5037
|
+
formatSettings = csvSettings;
|
|
5030
5038
|
// <- TODO: [🤹♂️] More universal, make simmilar pattern for other formats for example \n vs \r\n in text
|
|
5031
5039
|
}
|
|
5032
5040
|
return [4 /*yield*/, subvalueDefinition.mapValues(parameterValue, template.foreach.outputSubparameterName, formatSettings, function (subparameters, index) { return __awaiter(_this, void 0, void 0, function () {
|
|
@@ -5179,13 +5187,12 @@ function getReservedParametersForTemplate(options) {
|
|
|
5179
5187
|
*/
|
|
5180
5188
|
function executeTemplate(options) {
|
|
5181
5189
|
return __awaiter(this, void 0, void 0, function () {
|
|
5182
|
-
var currentTemplate, preparedPipeline, parametersToPass, tools, onProgress,
|
|
5183
|
-
var e_1,
|
|
5184
|
-
return __generator(this, function (
|
|
5185
|
-
switch (
|
|
5190
|
+
var currentTemplate, preparedPipeline, parametersToPass, tools, onProgress, $executionReport, pipelineIdentification, _a, maxExecutionAttempts, name, title, priority, usedParameterNames, dependentParameterNames, definedParameters, _b, _c, _d, definedParameterNames, parameters, _loop_1, _e, _f, parameterName, maxAttempts, jokerParameterNames, preparedContent, resultString;
|
|
5191
|
+
var e_1, _g, _h;
|
|
5192
|
+
return __generator(this, function (_j) {
|
|
5193
|
+
switch (_j.label) {
|
|
5186
5194
|
case 0:
|
|
5187
|
-
currentTemplate = options.currentTemplate, preparedPipeline = options.preparedPipeline, parametersToPass = options.parametersToPass, tools = options.tools, onProgress = options.onProgress,
|
|
5188
|
-
maxExecutionAttempts = settings.maxExecutionAttempts;
|
|
5195
|
+
currentTemplate = options.currentTemplate, preparedPipeline = options.preparedPipeline, parametersToPass = options.parametersToPass, tools = options.tools, onProgress = options.onProgress, $executionReport = options.$executionReport, pipelineIdentification = options.pipelineIdentification, _a = options.maxExecutionAttempts, maxExecutionAttempts = _a === void 0 ? DEFAULT_MAX_EXECUTION_ATTEMPTS : _a;
|
|
5189
5196
|
name = "pipeline-executor-frame-".concat(currentTemplate.name);
|
|
5190
5197
|
title = currentTemplate.title;
|
|
5191
5198
|
priority = preparedPipeline.templates.length - preparedPipeline.templates.indexOf(currentTemplate);
|
|
@@ -5200,7 +5207,7 @@ function executeTemplate(options) {
|
|
|
5200
5207
|
// <- [🍸]
|
|
5201
5208
|
})];
|
|
5202
5209
|
case 1:
|
|
5203
|
-
|
|
5210
|
+
_j.sent();
|
|
5204
5211
|
usedParameterNames = extractParameterNamesFromTemplate(currentTemplate);
|
|
5205
5212
|
dependentParameterNames = new Set(currentTemplate.dependentParameterNames);
|
|
5206
5213
|
// TODO: [👩🏾🤝👩🏻] Use here `mapAvailableToExpectedParameters`
|
|
@@ -5211,15 +5218,15 @@ function executeTemplate(options) {
|
|
|
5211
5218
|
.map(function (name) { return "{".concat(name, "}"); })
|
|
5212
5219
|
.join(', '), "\n\n ").concat(block(pipelineIdentification), "\n\n "); }));
|
|
5213
5220
|
}
|
|
5214
|
-
|
|
5215
|
-
|
|
5221
|
+
_c = (_b = Object).freeze;
|
|
5222
|
+
_d = [{}];
|
|
5216
5223
|
return [4 /*yield*/, getReservedParametersForTemplate({
|
|
5217
5224
|
preparedPipeline: preparedPipeline,
|
|
5218
5225
|
template: currentTemplate,
|
|
5219
5226
|
pipelineIdentification: pipelineIdentification,
|
|
5220
5227
|
})];
|
|
5221
5228
|
case 2:
|
|
5222
|
-
definedParameters =
|
|
5229
|
+
definedParameters = _c.apply(_b, [__assign.apply(void 0, [__assign.apply(void 0, _d.concat([(_j.sent())])), parametersToPass])]);
|
|
5223
5230
|
definedParameterNames = new Set(Object.keys(definedParameters));
|
|
5224
5231
|
parameters = {};
|
|
5225
5232
|
_loop_1 = function (parameterName) {
|
|
@@ -5239,15 +5246,15 @@ function executeTemplate(options) {
|
|
|
5239
5246
|
try {
|
|
5240
5247
|
// Note: [2] Check that all used parameters are defined and removing unused parameters for this template
|
|
5241
5248
|
// TODO: [👩🏾🤝👩🏻] Use here `mapAvailableToExpectedParameters`
|
|
5242
|
-
for (
|
|
5243
|
-
parameterName =
|
|
5249
|
+
for (_e = __values(Array.from(union(definedParameterNames, usedParameterNames, dependentParameterNames))), _f = _e.next(); !_f.done; _f = _e.next()) {
|
|
5250
|
+
parameterName = _f.value;
|
|
5244
5251
|
_loop_1(parameterName);
|
|
5245
5252
|
}
|
|
5246
5253
|
}
|
|
5247
5254
|
catch (e_1_1) { e_1 = { error: e_1_1 }; }
|
|
5248
5255
|
finally {
|
|
5249
5256
|
try {
|
|
5250
|
-
if (
|
|
5257
|
+
if (_f && !_f.done && (_g = _e.return)) _g.call(_e);
|
|
5251
5258
|
}
|
|
5252
5259
|
finally { if (e_1) throw e_1.error; }
|
|
5253
5260
|
}
|
|
@@ -5267,12 +5274,11 @@ function executeTemplate(options) {
|
|
|
5267
5274
|
template: currentTemplate,
|
|
5268
5275
|
preparedPipeline: preparedPipeline,
|
|
5269
5276
|
tools: tools,
|
|
5270
|
-
settings: settings,
|
|
5271
5277
|
$executionReport: $executionReport,
|
|
5272
5278
|
pipelineIdentification: pipelineIdentification,
|
|
5273
5279
|
})];
|
|
5274
5280
|
case 3:
|
|
5275
|
-
resultString =
|
|
5281
|
+
resultString = _j.sent();
|
|
5276
5282
|
return [4 /*yield*/, onProgress({
|
|
5277
5283
|
name: name,
|
|
5278
5284
|
title: title,
|
|
@@ -5284,12 +5290,12 @@ function executeTemplate(options) {
|
|
|
5284
5290
|
// <- [🍸]
|
|
5285
5291
|
})];
|
|
5286
5292
|
case 4:
|
|
5287
|
-
|
|
5288
|
-
return [2 /*return*/, Object.freeze((
|
|
5289
|
-
|
|
5293
|
+
_j.sent();
|
|
5294
|
+
return [2 /*return*/, Object.freeze((_h = {},
|
|
5295
|
+
_h[currentTemplate.resultingParameterName] =
|
|
5290
5296
|
// <- Note: [👩👩👧] No need to detect parameter collision here because pipeline checks logic consistency during construction
|
|
5291
5297
|
resultString,
|
|
5292
|
-
|
|
5298
|
+
_h))];
|
|
5293
5299
|
}
|
|
5294
5300
|
});
|
|
5295
5301
|
});
|
|
@@ -5348,13 +5354,12 @@ function filterJustOutputParameters(options) {
|
|
|
5348
5354
|
*/
|
|
5349
5355
|
function executePipeline(options) {
|
|
5350
5356
|
return __awaiter(this, void 0, void 0, function () {
|
|
5351
|
-
var inputParameters, tools, onProgress, pipeline, setPreparedPipeline, pipelineIdentification,
|
|
5357
|
+
var inputParameters, tools, onProgress, pipeline, setPreparedPipeline, pipelineIdentification, maxParallelCount, rootDirname, _a, isVerbose, preparedPipeline, errors, warnings, executionReport, isReturned, _b, _c, parameter, e_1_1, _loop_1, _d, _e, parameterName, state_1, e_2_1, parametersToPass, resovedParameterNames_1, unresovedTemplates_1, resolving_1, loopLimit, _loop_2, error_1, usage_1, outputParameters_1, usage, outputParameters;
|
|
5352
5358
|
var e_1, _f, e_2, _g;
|
|
5353
5359
|
return __generator(this, function (_h) {
|
|
5354
5360
|
switch (_h.label) {
|
|
5355
5361
|
case 0:
|
|
5356
|
-
inputParameters = options.inputParameters, tools = options.tools, onProgress = options.onProgress, pipeline = options.pipeline, setPreparedPipeline = options.setPreparedPipeline, pipelineIdentification = options.pipelineIdentification,
|
|
5357
|
-
maxParallelCount = settings.maxParallelCount, rootDirname = settings.rootDirname, _a = settings.isVerbose, isVerbose = _a === void 0 ? IS_VERBOSE : _a;
|
|
5362
|
+
inputParameters = options.inputParameters, tools = options.tools, onProgress = options.onProgress, pipeline = options.pipeline, setPreparedPipeline = options.setPreparedPipeline, pipelineIdentification = options.pipelineIdentification, maxParallelCount = options.maxParallelCount, rootDirname = options.rootDirname, _a = options.isVerbose, isVerbose = _a === void 0 ? DEFAULT_IS_VERBOSE : _a;
|
|
5358
5363
|
preparedPipeline = options.preparedPipeline;
|
|
5359
5364
|
if (!(preparedPipeline === undefined)) return [3 /*break*/, 2];
|
|
5360
5365
|
return [4 /*yield*/, preparePipeline(pipeline, tools, {
|
|
@@ -5539,12 +5544,7 @@ function executePipeline(options) {
|
|
|
5539
5544
|
return [3 /*break*/, 4];
|
|
5540
5545
|
case 3:
|
|
5541
5546
|
unresovedTemplates_1 = unresovedTemplates_1.filter(function (template) { return template !== currentTemplate; });
|
|
5542
|
-
work_1 = executeTemplate({
|
|
5543
|
-
currentTemplate: currentTemplate,
|
|
5544
|
-
preparedPipeline: preparedPipeline,
|
|
5545
|
-
parametersToPass: parametersToPass,
|
|
5546
|
-
tools: tools,
|
|
5547
|
-
onProgress: function (progress) {
|
|
5547
|
+
work_1 = executeTemplate(__assign(__assign({}, options), { currentTemplate: currentTemplate, preparedPipeline: preparedPipeline, parametersToPass: parametersToPass, tools: tools, onProgress: function (progress) {
|
|
5548
5548
|
if (isReturned) {
|
|
5549
5549
|
throw new UnexpectedError(spaceTrim(function (block) { return "\n Can not call `onProgress` after pipeline execution is finished\n\n ".concat(block(pipelineIdentification), "\n\n ").concat(block(JSON.stringify(progress, null, 4)
|
|
5550
5550
|
.split('\n')
|
|
@@ -5554,11 +5554,7 @@ function executePipeline(options) {
|
|
|
5554
5554
|
if (onProgress) {
|
|
5555
5555
|
onProgress(progress);
|
|
5556
5556
|
}
|
|
5557
|
-
},
|
|
5558
|
-
settings: settings,
|
|
5559
|
-
$executionReport: executionReport,
|
|
5560
|
-
pipelineIdentification: spaceTrim(function (block) { return "\n ".concat(block(pipelineIdentification), "\n Template name: ").concat(currentTemplate.name, "\n Template title: ").concat(currentTemplate.title, "\n "); }),
|
|
5561
|
-
})
|
|
5557
|
+
}, $executionReport: executionReport, pipelineIdentification: spaceTrim(function (block) { return "\n ".concat(block(pipelineIdentification), "\n Template name: ").concat(currentTemplate.name, "\n Template title: ").concat(currentTemplate.title, "\n "); }) }))
|
|
5562
5558
|
.then(function (newParametersToPass) {
|
|
5563
5559
|
parametersToPass = __assign(__assign({}, newParametersToPass), parametersToPass);
|
|
5564
5560
|
resovedParameterNames_1 = __spreadArray(__spreadArray([], __read(resovedParameterNames_1), false), [currentTemplate.resultingParameterName], false);
|
|
@@ -5662,8 +5658,7 @@ function executePipeline(options) {
|
|
|
5662
5658
|
*/
|
|
5663
5659
|
function createPipelineExecutor(options) {
|
|
5664
5660
|
var _this = this;
|
|
5665
|
-
var pipeline = options.pipeline, tools = options.tools, _a = options.
|
|
5666
|
-
var _b = settings.maxExecutionAttempts, maxExecutionAttempts = _b === void 0 ? MAX_EXECUTION_ATTEMPTS : _b, _c = settings.maxParallelCount, maxParallelCount = _c === void 0 ? MAX_PARALLEL_COUNT : _c, _d = settings.csvSettings, csvSettings = _d === void 0 ? DEFAULT_CSV_SETTINGS : _d, _e = settings.isVerbose, isVerbose = _e === void 0 ? IS_VERBOSE : _e, _f = settings.isNotPreparedWarningSupressed, isNotPreparedWarningSupressed = _f === void 0 ? false : _f, _g = settings.rootDirname, rootDirname = _g === void 0 ? null : _g;
|
|
5661
|
+
var pipeline = options.pipeline, tools = options.tools, _a = options.maxExecutionAttempts, maxExecutionAttempts = _a === void 0 ? DEFAULT_MAX_EXECUTION_ATTEMPTS : _a, _b = options.maxParallelCount, maxParallelCount = _b === void 0 ? DEFAULT_MAX_PARALLEL_COUNT : _b, _c = options.csvSettings, csvSettings = _c === void 0 ? DEFAULT_CSV_SETTINGS : _c, _d = options.isVerbose, isVerbose = _d === void 0 ? DEFAULT_IS_VERBOSE : _d, _e = options.isNotPreparedWarningSupressed, isNotPreparedWarningSupressed = _e === void 0 ? false : _e, _f = options.rootDirname, rootDirname = _f === void 0 ? null : _f;
|
|
5667
5662
|
validatePipeline(pipeline);
|
|
5668
5663
|
var pipelineIdentification = (function () {
|
|
5669
5664
|
// Note: This is a 😐 implementation of [🚞]
|
|
@@ -5697,14 +5692,12 @@ function createPipelineExecutor(options) {
|
|
|
5697
5692
|
tools: tools,
|
|
5698
5693
|
onProgress: onProgress,
|
|
5699
5694
|
pipelineIdentification: spaceTrim(function (block) { return "\n ".concat(block(pipelineIdentification), "\n ").concat(runCount === 1 ? '' : "Run #".concat(runCount), "\n "); }),
|
|
5700
|
-
|
|
5701
|
-
|
|
5702
|
-
|
|
5703
|
-
|
|
5704
|
-
|
|
5705
|
-
|
|
5706
|
-
rootDirname: rootDirname,
|
|
5707
|
-
},
|
|
5695
|
+
maxExecutionAttempts: maxExecutionAttempts,
|
|
5696
|
+
maxParallelCount: maxParallelCount,
|
|
5697
|
+
csvSettings: csvSettings,
|
|
5698
|
+
isVerbose: isVerbose,
|
|
5699
|
+
isNotPreparedWarningSupressed: isNotPreparedWarningSupressed,
|
|
5700
|
+
rootDirname: rootDirname,
|
|
5708
5701
|
})];
|
|
5709
5702
|
});
|
|
5710
5703
|
}); };
|
|
@@ -5726,7 +5719,7 @@ var markdownScraperMetadata = $deepFreeze({
|
|
|
5726
5719
|
mimeTypes: ['text/markdown', 'text/plain'],
|
|
5727
5720
|
documentationUrl: 'https://github.com/webgptorg/promptbook/discussions/@@',
|
|
5728
5721
|
isAvilableInBrowser: true,
|
|
5729
|
-
requiredExecutables: [
|
|
5722
|
+
requiredExecutables: [],
|
|
5730
5723
|
}); /* <- TODO: [🤛] */
|
|
5731
5724
|
/**
|
|
5732
5725
|
* Registration of known scraper metadata
|
|
@@ -5770,7 +5763,7 @@ var MarkdownScraper = /** @class */ (function () {
|
|
|
5770
5763
|
return __generator(this, function (_k) {
|
|
5771
5764
|
switch (_k.label) {
|
|
5772
5765
|
case 0:
|
|
5773
|
-
_a = this.options, _b = _a.maxParallelCount, maxParallelCount = _b === void 0 ?
|
|
5766
|
+
_a = this.options, _b = _a.maxParallelCount, maxParallelCount = _b === void 0 ? DEFAULT_MAX_PARALLEL_COUNT : _b, _c = _a.isVerbose, isVerbose = _c === void 0 ? DEFAULT_IS_VERBOSE : _c;
|
|
5774
5767
|
llm = this.tools.llm;
|
|
5775
5768
|
if (llm === undefined) {
|
|
5776
5769
|
throw new MissingToolsError('LLM tools are required for scraping external files');
|
|
@@ -5869,7 +5862,8 @@ var MarkdownScraper = /** @class */ (function () {
|
|
|
5869
5862
|
embeddingResult = _c.sent();
|
|
5870
5863
|
index.push({
|
|
5871
5864
|
modelName: embeddingResult.modelName,
|
|
5872
|
-
position: embeddingResult.content,
|
|
5865
|
+
position: __spreadArray([], __read(embeddingResult.content), false),
|
|
5866
|
+
// <- TODO: [🪓] Here should be no need for spreading new array, just `position: embeddingResult.content`
|
|
5873
5867
|
});
|
|
5874
5868
|
_c.label = 6;
|
|
5875
5869
|
case 6: return [3 /*break*/, 8];
|
|
@@ -5920,7 +5914,7 @@ var documentScraperMetadata = $deepFreeze({
|
|
|
5920
5914
|
mimeTypes: ['application/vnd.openxmlformats-officedocument.wordprocessingml.document'],
|
|
5921
5915
|
documentationUrl: 'https://github.com/webgptorg/promptbook/discussions/@@',
|
|
5922
5916
|
isAvilableInBrowser: false,
|
|
5923
|
-
requiredExecutables: ['
|
|
5917
|
+
requiredExecutables: ['Pandoc'],
|
|
5924
5918
|
}); /* <- TODO: [🤛] */
|
|
5925
5919
|
/**
|
|
5926
5920
|
* Registration of known scraper metadata
|
|
@@ -5962,11 +5956,11 @@ var DocumentScraper = /** @class */ (function () {
|
|
|
5962
5956
|
DocumentScraper.prototype.$convert = function (source) {
|
|
5963
5957
|
var _a;
|
|
5964
5958
|
return __awaiter(this, void 0, void 0, function () {
|
|
5965
|
-
var _b, _c, rootDirname, _d, cacheDirname, _e,
|
|
5959
|
+
var _b, _c, rootDirname, _d, cacheDirname, _e, intermediateFilesStrategy, _f, isVerbose, extension, cacheFilehandler, command_1;
|
|
5966
5960
|
return __generator(this, function (_g) {
|
|
5967
5961
|
switch (_g.label) {
|
|
5968
5962
|
case 0:
|
|
5969
|
-
_b = this.options, _c = _b.rootDirname, rootDirname = _c === void 0 ? process.cwd() : _c, _d = _b.cacheDirname, cacheDirname = _d === void 0 ?
|
|
5963
|
+
_b = this.options, _c = _b.rootDirname, rootDirname = _c === void 0 ? process.cwd() : _c, _d = _b.cacheDirname, cacheDirname = _d === void 0 ? DEFAULT_SCRAPE_CACHE_DIRNAME : _d, _e = _b.intermediateFilesStrategy, intermediateFilesStrategy = _e === void 0 ? DEFAULT_INTERMEDIATE_FILES_STRATEGY : _e, _f = _b.isVerbose, isVerbose = _f === void 0 ? DEFAULT_IS_VERBOSE : _f;
|
|
5970
5964
|
if (!$isRunningInNode()) {
|
|
5971
5965
|
throw new KnowledgeScrapeError('Scraping .docx files is only supported in Node environment');
|
|
5972
5966
|
}
|
|
@@ -5985,7 +5979,7 @@ var DocumentScraper = /** @class */ (function () {
|
|
|
5985
5979
|
return [4 /*yield*/, getScraperIntermediateSource(source, {
|
|
5986
5980
|
rootDirname: rootDirname,
|
|
5987
5981
|
cacheDirname: cacheDirname,
|
|
5988
|
-
|
|
5982
|
+
intermediateFilesStrategy: intermediateFilesStrategy,
|
|
5989
5983
|
extension: 'md',
|
|
5990
5984
|
isVerbose: isVerbose,
|
|
5991
5985
|
})];
|
|
@@ -5995,10 +5989,8 @@ var DocumentScraper = /** @class */ (function () {
|
|
|
5995
5989
|
case 2:
|
|
5996
5990
|
if (!!(_g.sent())) return [3 /*break*/, 5];
|
|
5997
5991
|
command_1 = "\"".concat(this.tools.executables.pandocPath, "\" -f ").concat(extension, " -t markdown \"").concat(source.filename, "\" -o \"").concat(cacheFilehandler.filename, "\"");
|
|
5998
|
-
// TODO: !!!!!! [🕊] Make execCommand standard (?node-)util of the promptbook
|
|
5999
5992
|
return [4 /*yield*/, $execCommand(command_1)];
|
|
6000
5993
|
case 3:
|
|
6001
|
-
// TODO: !!!!!! [🕊] Make execCommand standard (?node-)util of the promptbook
|
|
6002
5994
|
_g.sent();
|
|
6003
5995
|
return [4 /*yield*/, isFileExisting(cacheFilehandler.filename, this.tools.fs)];
|
|
6004
5996
|
case 4:
|
|
@@ -6077,7 +6069,11 @@ var legacyDocumentScraperMetadata = $deepFreeze({
|
|
|
6077
6069
|
mimeTypes: ['application/msword', 'text/rtf'],
|
|
6078
6070
|
documentationUrl: 'https://github.com/webgptorg/promptbook/discussions/@@',
|
|
6079
6071
|
isAvilableInBrowser: false,
|
|
6080
|
-
requiredExecutables: [
|
|
6072
|
+
requiredExecutables: [
|
|
6073
|
+
'Pandoc',
|
|
6074
|
+
'LibreOffice',
|
|
6075
|
+
// <- TODO: [🧠] Should be 'LibreOffice' here, its dependency of dependency
|
|
6076
|
+
],
|
|
6081
6077
|
}); /* <- TODO: [🤛] */
|
|
6082
6078
|
/**
|
|
6083
6079
|
* Registration of known scraper metadata
|
|
@@ -6119,11 +6115,11 @@ var LegacyDocumentScraper = /** @class */ (function () {
|
|
|
6119
6115
|
LegacyDocumentScraper.prototype.$convert = function (source) {
|
|
6120
6116
|
var _a;
|
|
6121
6117
|
return __awaiter(this, void 0, void 0, function () {
|
|
6122
|
-
var _b, _c, rootDirname, _d, cacheDirname, _e,
|
|
6118
|
+
var _b, _c, rootDirname, _d, cacheDirname, _e, intermediateFilesStrategy, _f, isVerbose, extension, cacheFilehandler, documentSourceOutdirPathForLibreOffice_1, command_1, files_1, file;
|
|
6123
6119
|
return __generator(this, function (_g) {
|
|
6124
6120
|
switch (_g.label) {
|
|
6125
6121
|
case 0:
|
|
6126
|
-
_b = this.options, _c = _b.rootDirname, rootDirname = _c === void 0 ? process.cwd() : _c, _d = _b.cacheDirname, cacheDirname = _d === void 0 ?
|
|
6122
|
+
_b = this.options, _c = _b.rootDirname, rootDirname = _c === void 0 ? process.cwd() : _c, _d = _b.cacheDirname, cacheDirname = _d === void 0 ? DEFAULT_SCRAPE_CACHE_DIRNAME : _d, _e = _b.intermediateFilesStrategy, intermediateFilesStrategy = _e === void 0 ? DEFAULT_INTERMEDIATE_FILES_STRATEGY : _e, _f = _b.isVerbose, isVerbose = _f === void 0 ? DEFAULT_IS_VERBOSE : _f;
|
|
6127
6123
|
if (!$isRunningInNode()) {
|
|
6128
6124
|
throw new KnowledgeScrapeError('Scraping .doc files is only supported in Node environment');
|
|
6129
6125
|
}
|
|
@@ -6142,7 +6138,7 @@ var LegacyDocumentScraper = /** @class */ (function () {
|
|
|
6142
6138
|
return [4 /*yield*/, getScraperIntermediateSource(source, {
|
|
6143
6139
|
rootDirname: rootDirname,
|
|
6144
6140
|
cacheDirname: cacheDirname,
|
|
6145
|
-
|
|
6141
|
+
intermediateFilesStrategy: intermediateFilesStrategy,
|
|
6146
6142
|
extension: 'docx',
|
|
6147
6143
|
isVerbose: isVerbose,
|
|
6148
6144
|
})];
|
|
@@ -6158,10 +6154,8 @@ var LegacyDocumentScraper = /** @class */ (function () {
|
|
|
6158
6154
|
.split('\\')
|
|
6159
6155
|
.join('/');
|
|
6160
6156
|
command_1 = "\"".concat(this.tools.executables.libreOfficePath, "\" --headless --convert-to docx \"").concat(source.filename, "\" --outdir \"").concat(documentSourceOutdirPathForLibreOffice_1, "\"");
|
|
6161
|
-
// TODO: !!!!!! [🕊] Make execCommand standard (?node-)util of the promptbook - this should trigger build polution error
|
|
6162
6157
|
return [4 /*yield*/, $execCommand(command_1)];
|
|
6163
6158
|
case 3:
|
|
6164
|
-
// TODO: !!!!!! [🕊] Make execCommand standard (?node-)util of the promptbook - this should trigger build polution error
|
|
6165
6159
|
_g.sent();
|
|
6166
6160
|
return [4 /*yield*/, readdir(documentSourceOutdirPathForLibreOffice_1)];
|
|
6167
6161
|
case 4:
|